The bulk electrical polarization P of a material in an electric field (or the electric part of an optical field) may be expanded in powers of the field according to equal, where P.sub.S is the spontaneous polarization (i.e. polarization present in the absence of applied field, X.sup.(1) is the linear polarizability, X.sup.(2) is the second order nonlinear hyperpolarizability, or second order nonlinear susceptibility, X.sup.(3) is the third order nonlinear hyperpolarizability or third order nonlinear susceptibility. The subscripts i, j, k etc. correspond to the Cartesian coordinates x, y, or z for the system (Williams, D. J., (1984) Angew. Chem. Int. Ed. Engl. 23:690-703). ##EQU1##
The sum of all terms to the right of P.sub.S in equal give the induced bulk polarization in response to an applied field or fields. The spontaneous polarization P.sub.S is a vector, while the susceptibilities X.sup.(1) etc. are tensors with component values which are dependent upon the frequency of the applied fields. The square of X.sup.(1) driven by a DC or low frequency AC field is proportional to the dielectric constant of the material, while the square of X.sup.(1) driven by an optical frequency AC field is proportional to the refractive index of the material. All materials possess non-zero X.sup.(1) and X.sup.(3). There are certain symmetry requirements for P.sub.s and for .sup.(2)), however. Thus, in order to possess non-zero P.sub.s, the system must have polar symmetry. Furthermore, within the electronic dipolar model, X.sup.(2) is zero unless the system possesses noncentrosymmetric symmetry (acentric). All materials with polar symmetry are acentric, but not all acentric materials are polar. Thus it is possible for a material to possess strictly zero P.sub.s by symmetry, but non-zero X.sup.(2) in the electronic dipolar model.
Materials possessing non-zero X.sup.(2) exhibit many effects of great current and potential utility. These include but are not limited to: 1) Second harmonic generation (SHG); 2) Sum and difference frequency generation; 3) Optical parametric amplification; 4) Optical rectification; and 5) A linear electrooptic effect (Pockel's effect). Effects 1, 2 and 3 depend upon the induction of optical frequency AC polarizations (or charge flow in the material changing in sign or magnitude at optical frequencies) in the material in response to optical frequency AC applied fields, and therefore derive from optical frequency X.sup.(2) values. These values of X.sup.(2) may be termed "ultrafast".
In general, the ultrafast X.sup.(2) is a lower limit, and the induced polarization in response to lower frequency applied fields will in general be larger (i.e. X.sup.(2) generally increases with decreasing driving field frequency, though the increase is not monotonic). Very large increases in X.sup.(2) occur at frequencies where resonant absorption of the driving radiation occurs. For the applications of interest in this invention, however, non-resonant interactions of the material with driving and induced fields are preferred.
Currently X.sup.(2) materials are utilized extensively for frequency conversion (effects 1 and 2 above), and more experimentally in electro-optic modulators (effect 5). Typically these materials are inorganic single crystals (for example single crystals of potassium dihydrogen phosphate (KDP) or lithium niobate (LiNbO.sub.3). For many applications, particularly in the emerging opto-electronics and photonics industry, easily processible thin films possessing X.sup.(2) are of great potential utility. Uses of X.sup.(2) thin films include, for example, electro-optic switching and frequency processing in guided-wave geometries. Guided-wave geometries are useful in, for example, integrated optical circuits or specialized devices such as optical parametric amplifiers or electro-optic modulators.
For some of these potential applications, the film must work in concert with other materials, such as silicon or other semiconductor integrated circuits. This requires that the film be processed onto or with the semiconductor or other material in a controlled way, affording a hybrid device. In some thin film applications inorganic crystals are relatively difficult to utilize, being difficult to hybridize with semiconductors.
It has been known for some time that organic materials possess potential advantages in X.sup.(2) applications (see Prasad, P. N., (1990) Chem. Mater. 2(6):660-669). These include: 1) Easy processibility relative to inorganic crystals; and 2) Potentially large and relatively easily tuned values of the X.sup.(2) components. The potential for easy processibility derives by analogy to the relatively easy creation of, for example, organic polymer and liquid crystal films of optical quality. The potential for large X.sup.(2) derives in part from experimentally determined values of the molecular susceptibilities of organic molecules. Thus, the polarization of a molecule in the presence of applied electric fields is given by equ 2, where .mu. is the molecular dipole moment, .alpha. is the molecular linear polarizability, .beta. is the molecular second order hyperpolarizability, etc. ##EQU2##
Using the technique of, for example, electric field induced second harmonic generation (EFISH) in isotropic solutions, it is possible to measure the magnitude of certain components of .beta. for many organic molecules. A fairly good estimate of X.sup.(2) may be made based upon these .beta. values, and using such estimates, it may be shown that X.sup.(2) for organic materials may in principle be much larger than those exhibited by inorganic crystals (Prasad, P. N., (1990) supra).
Finally, the potential for tunability derives from the great structural diversity of organic molecules combined with some relatively simple models for the molecular origins of .beta.. Thus, while the level of understanding of the molecular origins of .beta. is not quantitative, it is easy to predict-qualitatively the magnitude of .beta. expected for new organic molecules using the "two-state model" (Williams, D. J., (1984) supra and Prasad, P. N., (1990) supra). In this model, using the valence bond structures of the ground and first electronically excited states one may calculate the expected .beta. value in response to driving fields far from resonance. At the current state of the art, these calculations give only a qualitative picture of the hyperpolarizability of the molecules.
More simply, it is generally appreciated that the molecular .beta. increases with increasing difference in molecular dipole moment of the ground state and first electronically excited state of the molecule. Furthermore it is generally known that this occurs when a donor group and an acceptor group are oriented ortho or para on a benzene ring. Donor and acceptor refer to the ability of the group to either donate electron density (donor) or donate positive charge density (acceptor) to an aromatic ring. When the donor and acceptor are regiochemically placed on the benzene ring such that negative charge transfer from the donor to the acceptor can occur according to simple resonance arguments (i.e. they are conjugated), then a large .beta. should result. With a single benzene ring the conjugated substituents are ortho or para, as indicated in the following diagram for p-nitroaniline--a prototypical organic NLO molecule. An axis, termed here the ".beta. axis", may be defined for such molecules. This axis is along the line connecting the donor substituent with the acceptor substituent. If the molecular coordinate system is defined such that the .beta. axis is along y, then .beta..sub.y,yy will be a large component of the .beta. tensor. ##STR2## Furthermore, it is understood that the larger the distance between the donor and acceptor, the larger the .beta. which will result for a given donor-acceptor pair (.beta. goes up approximately as the square of the distance separating the donor and acceptor). Thus, para nitroaniline has a larger .beta. than the ortho isomer. Furthermore, conjugation between the donor and acceptor can be across a larger grouping than one benzene ring, as long as the conjugation is not broken. Thus, stilbenes, tolanes, and diphenyl azo compounds substituted at the p-p' positions with donor and acceptor groups possess very large .beta. values, with the .beta. axis on the line between the donor and acceptor, stilbenes and azo compounds larger than tolanes. The prototypical organic molecule with very large .beta. is disperse red 1, a diphenyl azo compound, whose ground state structure and charge transfer resonance structure are: ##STR3##
Furthermore, it is known that when .beta. gets large with increasing conjugation, then the farthest red resonant electronic absorption peak (.gamma..sub.max) is red-shifted, leading to increased resonant absorption at longer wavelengths close to .gamma..sub.max relative to a molecule with less conjugation or a smaller conjugation length. For some applications blue-shifted resonant absorption (i.e. towards the UV, affording more visible clarity) is advantageous, such as frequency doubling into the blue part of the visible spectrum. For some applications, such as electro-optic modulators which typically operate at wavelengths where glass fiber has the minimum dispersion (&gt;1.0 .mu.m) the red shift in .gamma..sub.max for the NLO molecules may not be a disadvantage.
From the above discussion and data it can be seen that organic molecules may be designed qualitatively to possess a given .beta. value consistent with a given .gamma..sub.max. It is not the subject of this invention to teach new donor-acceptor pairs or new conjugating spacers. Rather, this invention can take advantage of most known or new donor-acceptor pairs, and also known or new conjugating spacers.
It is known that in order to possess useful X.sup.(2) the NLO molecules must be combined to create a material, in some cases a thin film, and in some cases a bulk sample, typically much larger than the size of the molecules, but possessing acentric symmetry (Williams, D. J., (1984) supra and Prasad, P. N. (1990) supra). Furthermore, it is sufficient but not necessary for the material to possess polar order. Furthermore, it is understood that when the donor, acceptor and conjugating spacer (the B axis) lie on or close to a polar axis of a medium with polar order, and are oriented along the polar axis in a polar fashion, then the X.sup.(2) of the material is optimized for that donor-conjugating spacer-acceptor unit.
Several methods for achieving the combination of the NLO molecules into the desired X.sup.(2) material are known. These include: 1) Single crystals or oriented microcrystalline solids (see for example Marder, S. R., et al., (1989) Science 245:626-628); 2) Langmuir-Blodgett multilayers or self-assembling multilayers (see for example Popovitz-Biro, R., et al., (1988) J. Am. Chem. Soc. 110(8):2672-2674) and Tillman, N., et al., (1988) J. Am. Chem. Soc. 110:6136-6144); and 3) Electrically poled polymer films (see for example Williams, D. J. (1984) supra, Dembek, A. A., et al., (1990) Chemistry of Materials 2(2):97-99, and Park, J., et al., (1990) Chem. Mater. 2:229-231). The present invention provides a new method for achieving the NLO material by combining donor-conjugating spacer-acceptor arrays (the .beta. axis) oriented with good polar order along the polar axis of ferroelectric liquid crystal samples, which are typically but not necessarily thin films. Furthermore, the method of the present invention possesses important advantages over any of the previously existing methods.
It is known that often the molecular .beta. impart a large molecular dipole moment in the ground state along the .beta. axis, and that upon crystallization, these units often orient antiparallel to afford centrosymmetric symmetry in the crystal. Thus, for example, p-nitroaniline, while possessing a useful .beta. value, gives centrosymmetric crystals with very small or zero X.sup.(2) (exactly zero in the electronic dipolar model).
It is also known that it is often possible by relatively small modifications to the structure of the molecules, to obtain polar crystals with good polar orientation of the .beta. axis along the polar axis. Thus, for example, when the p-nitroaniline molecule is substituted with a methyl group ortho to the nitro grouping, the resulting methyl nitroaniline (MNA) fortuitously crystallizes with appropriate symmetry for X.sup.(2), and indeed MNA crystals have a very large X.sup.(2) value with moderate resonant visible absorption.
It is also known that organic single crystals, especially those formed from non-ionic molecules, often called Van der Waals crystals, are typically difficult to process into optical quality materials. This may be due to the fact that crystal growth is a kinetic, rather than a thermodynamic phenomenon, and crystal nucleation at multiple sites leads to the formation of domain wall defects which scatter light. The scattering of light from defects is highly undesirable in NLO materials. Furthermore, it is also difficult to control the organic crystal growth in thin film applications, especially for hybrid devices where the organic film must be oriented correctly on a specific substrate surface.
This lack of processibility presents a major disadvantage of organic crystals for X.sup.(2) applications. The problems with crystals led to the invention of several alternative approaches, especially for thin film applications. Two such approaches involve growing a crystalline or non-crystalline film from a substrate one molecular monolayer at a time. In the Langmuir-Blodgett method, NLO molecules are synthesized such that they also form monolayers (LB films) on water. This is achieved by controlling the hydrophilicity and hydrophoebicityof parts of the molecules, and can lead to excellent control over the orientation of molecular fragments, including the functional arrays affording large .beta., in the monolayers on water.
By careful dipping of a substrate into the monolayer film it is possible to deposit the monolayer with good structural control onto the surface of the substrate. Additional dipping cycles, with additional methods needed to achieve bulk polar order, can afford multilayers with appropriate structure for X.sup.(2). In a somewhat related process, it is possible with correctly designed NLO molecules possessing reactive functional groupings to dip a substrate into an isotropic solution of NLO molecules, and obtain a structurally well-defined monolayer covalently bound to the substrate. Chemical modification of the resulting new surface to introduce appropriate reactivity at the surface, followed by another dipping cycle, etc., can afford multilayers with structure appropriate for NLO applications.
In both of these approaches, many dipping cycles (&gt;1,000) are required to achieve materials of good utility for NLO applications. Furthermore, the structural stability of the resulting multilayers, especially the LB multilayers, is not currently known. In addition, the optical quality achievable for such films is not known. Finally, these techniques are inefficient in time, and presumably cost, and limit the possible geometries of the .beta. axis relative to the substrate surface since for all known examples the polar axis must be normal to the substrate surface. These factors represent disadvantages of the multilayer approaches for creation of X.sup.(2) thin films.
Molecules possessing a dipole moment and .beta. axis (typically colinear) can be either doped into a polymer matrix, or covalently attached to the polymer matrix. When such polymer matrix is then heated to a temperature above a glass transition temperature and subjected to an electric field, the NLO molecules will tend to align with their dipoles parallel to the field, affording the bulk polar order giving X.sup.(2). If the field is removed, the NLO molecules rapidly revert to their random state, destroying the X.sup.(2) of the sample. However, if the sample is cooled with the field applied below the polymer glass transition, then the polar order induced by the field can be "frozen" into the sample. The field can then be removed to give an optical quality film with useful X.sup.(2).
It is known that such films are typically unstable over time. That is, the NLO molecules, whether covalently attached to the polymer molecules, or doped into the polymer, will over time randomize their orientation, destroying some of the X.sup.(2) of the sample. While many approaches for stabilizing the polar order present in such films are being explored (chiefly cross-linking of the polymer lattice to stabilize the positions of the molecules temporally), the polar order in such films in the absence of applied fields is inherently unstable thermodynamically. In addition, the degree of orientation of the molecular .beta. axis along the polar axis achievable with the largest possible poling electric fields is relatively poor. The good optical quality of non-crystalline or microcrystalline polymer films, and their relative ease of manufacture, are advantages of the poled polymer method for creating X.sup.(2) films. The thermodynamic instability of X.sup.(2) poled polymers and the poor degree of structural control in the films are disadvantages of the method.
In liquid crystals (LCs), mesogen molecules spontaneously self-assemble into true fluids which are anisotropic. Typically, the mesogens are rod-shapped molecules. The long axis of the molecules, and also the optic axis of the LC phase, is called the director, which is represented by the unit vector n. It is relevant for the present invention that for all known LC phases, all properties of the phases are invariant with sign of the director (often represented as n.fwdarw.n). Thus, there is no spontaneous polar order along the director for any known LC phases.
When the liquid crystal is such that the molecules self-assemble into a layered structure, the liquid crystal phase is called smectic. Smectic LCs may be considered as a stack of 2-dimensional fluid phases each approximately one molecular length in thickness. There are many smectic LC structures. In some of these, the director is tilted coherently with respect to the layer normal (z), affording a tilted, layered structure. In this case, the thickness of the molecular layers is typically smaller than the molecular length. The plane containing n and z is termed the tilt plane.
While there is no fundamental reason why smectic C phases cannot possess spontaneous polar order, to our knowledge no smectic C phase possessing such order has ever been reported. Thus, all known smectic C phases possess the following symmetry elements for the phase: 1) A C.sub.2 axis of symmetry normal to the tilt plane (satisfying the empirical fact that n.fwdarw.nn); and 2) A .sigma. (mirror) plane congruent with the tilt plane. Thus, known smectic C phases possess a center of symmetry, i.e. they are centrosymmetric, and therefore possess zero X.sup.(2) in the electronic dipolar model.
When a medium is composed of chirally asymmetric molecules, such medium must be acentric, since the medium cannot possess any reflection symmetry. This is true for all media, including specifically isotropic liquids, all LC phases, and all crystalline or amorphous materials (Giordmaine, J. A., (1965) Phys. Rev. 138(6A):A1599-A1606, Rentzepis, P. M., et al., (1966) Phys. Rev. Lett. 16(18):792-794). The chirality does not, however, force polar order on the system.
For example, it has been demonstrated that chiral, isotropic liquids such as solutions of sugar molecules in water possess non-zero X.sup.(2) due to the acentricity of the medium (Rentzepis, P. M. (1966) supra). In such isotropic liquids there is no polar order, and thus no possibility for orientation of a molecular .beta. axis along a polar axis. In general, it is known that orientation of a large .beta. axis along a polar axis is a valid method for achieving large X.sup.(2). It is known that the X.sup.(2) occurring in acentric isotropic liquids is small.
Furthermore, chiral molecules possessing large .beta. are often utilized for growth of crystals for X.sup.(2) applications. Such crystals must be acentric, and may or may not possess polar order. However, even when polar order exists, in order to achieve large X.sup.(2) it is generally appreciated that the .beta. axis should be oriented along the polar axis. If the .beta. axis is not oriented along the polar axis, small X.sup.(2) will result.
When molecules in the smectic C or any other tilted smectic phase are made chirally asymmetric, then by symmetry considerations the phase must possess polar order in addition to acentricity. That is, all chiral fluid media (and non-fluid media) are acentric, but for known fluids, only in the tilted, layered LC case does the chirality also impart polar order upon the system (Walba, D. M. (1991) Ferroelectric Liquid Crystals: A Unique State of Matter. In: Mallouk T. E., ed. Advances in the Synthesis and Reactivity of Solids, Vol 1. Greenwich, Conn.: JAI Press Ltd 173-235). In the case of the smectic C phase, such a chiral smectic C phase is denoted as the smectic C, phase, which must possess polar order, its symmetry elements being limited to one C.sub.2 axis of symmetry, congruent with the polar axis of the phase, and oriented normal to the tilt plane. Such chiral, tilted, layered LCs are the only known fluids possessing thermodynamically stable polar order.
Typically, the polar order occurring smectic C* phases causes the spontaneous formation of a macroscopic electric dipole moment for the phase. The direction of this macroscopic dipole moment switches upon application of an external electric field, though external fields are not required for the macroscopic dipole to exist. Chiral smectic C, phases, and other chiral tilted, layered LC phases, are thus typically ferroelectric, and are often termed ferroelectric liquid crystals (FLCs) (Walba, D. M. (1991) supra). It is understood that this term includes all chiral, tilted, layered LC phases.
The macroscopic dipole moment of the phase present in the absence of applied electric fields is termed the ferroelectric polarization, P, which is the same as P.sub.s in equation 1. This polarization derives from the orientation of molecular dipoles (.mu. in equation 2) along the polar axis of the phase. The polarization P has a sign, which by arbitrary convention is positive when P (from negative to positive poles) points along the unit vector z .times.n, and negative when P is opposed to z.times.n. Enantiomeric (i.e. mirror image) FLC phases possess exactly equal magnitude but opposite sign of P (Walba, D. M. (1991) supra).
The experimental fact that FLCs possess polar order means that FLCs must possess non-zero X.sup.(2) in the electronic dipolar model. In addition, may FLC mesogens, including, for example, DOBAMBC, the first FLC ever reported, also possess functional arrays expected to have large .beta.. Thusr DOBAMBC and many other FLCs possess polar order and are composed of molecules with large .beta..
However, the measured values of the ultrafast X.sup.(2) in previously known FLCs are very small. Table 1 lists the values of X.sup.(2) for several exemplary known FLC materials as measured by the angle phase-matched SHG technique (see Taguchi, A., et al., (1989) Jpn. J. Appl. Phys. 28(6):L 997-L 999. and Liu, J. Y., et al., (1990) Optics Letters 15(5):267-269). Here, X.sup.(2) is given as value for the d-tensor coefficients (d=X.sup.(2) for SHG). For DOBAMBC and the commercial mixture ZLI 3654, only d.sub.eff is given. This value derives from a geometrical combination of various d coefficients, and the square of d.sub.eff is proportional to the intensity of second harmonic light output from the sample at the top on an angle phase-matched peak. Experiments providing the values of all non-zero components of the d tensor for the commercial mixture SCE 9 in the homeotropic alignment geometry have been accomplished, said values given in the table (Liu (1990) supra). Note that there is some correlation between the polarization of the sample and the observed d.sub.eff. This correlation can be indicative, but is not rigorous.
TABLE 1 __________________________________________________________________________ Values of the ferroelectric polarization, SHG efficiency, and X.sup.(2) (d.sub.eff and d coefficients), for representative previously known FLCs. Values for some common inorganic NLO crystals are included for comparison. SHG d Entry P arb d.sub.eff coefficients number compound (nC/cm.sup.2) units* (pm/V) (pm/V) __________________________________________________________________________ 1 DOBAMBC.sup.a -3 1 0.0008 2 ZLI 3654.sup.b -29 40 0.005 3 SCE 9.sup.c +33.6 160 0.01 d.sub.2,3 = 0.073 d.sub.2,2 = 0.027 d.sub.2,1 = 0.0026 d.sub.2,5 = 0.0009 4 KDP.sup.d d.sub.3,6 = 0.038 5 5% MgO:LiNbO.sub.3.sup.d d.sub.3,1 = -4.7 __________________________________________________________________________ *Intensity of the second harmonic light at the top of the type 1 eeo angl phasematched peak. .sup.a Vtyurin, A. N., et al., (1981) Phys. Status Solidi B 107(2):397-402. .sup.b Taguchi (1989) supra. .sup.c Liu (1990) supra. .sup.d Eckardt, R. C., et al., (1990) IEEE Journal of Quantum Electronics 26(5):922-933.
As can be seen from Table 1, the values of the largest d coefficients (one measure of X.sup.(2)) for the known FLCs which have been evaluated for X.sup.(2) are small relative to the known X.sup.(2) crystal KDP. This may be due to a combination of two factors: 1) The .beta. axis in FLCs is generally oriented along n, and there is no polar order along n; and 2) The degree of net polar order, as evidenced by the magnitude of the macroscopic polarization, is poor.
In this invention we provide a general approach for obtaining molecules which, when introduced into an FLC phase either as a pure mesogen, or component of an FLC mixture, will impart large X.sup.(2) to the FLC phase by orientation of a .beta. axis of the molecules along the polar axis in the FLC phase, and by achieving a high degree of polar order.
Typical thermotropic LC mesogens or components possess structures combining a rigid core with two relatively "floppy" tails (see Demus et at. (1974) Flussige Kristalle in Taballen, VEB Deutscher Verlag fur Grundstoffindustrie, Liebzig for a compilation of the molecular structures of LC molecules). FLC materials have been prepared by introduction of at least one stereocenter in at least of of the tails. Thus, referring to the general formula A, the rigid core can be, for example, benzylideneamino cinnamyl, biphenyl, phenylbenzoate, phenylpyrimidine or biphenylbenzoate, X and/or Y can be oxygen or CH.sub.2, R' is an achiral alkyl grouping with from five to twelve carbon atoms, and R* is a chiral moiety. ##STR4##
The FLCs reported to date are generally designed for use in the Clark-Lagerwaal surface stabilized FLC light valve (Clark, N. A., et al., (1980) Appl. Phys. Lett. 36(11):899-901), or other similar light modulation technology involving large nuclear motions of the FLC molecules for switching in response to applied DC fields or low frequency AC fields (&lt;100 MHz). For such FLCs, an important figure of merit is the characteristic response time of the cell (.tau.), given approximately by equ 3: ##EQU3## where .eta. is the orientational viscosity and P is the magnitude of the ferroelectric polarization density. The polarization typically derives from the type of chiral tail used, while the viscosity is a function of the core and chiral tail. The first FLC compound to be characterized was DOBAMBC, which contains a benzylideneaminocinnamyl core, a n-decyloxy achiral tail and 2-methylbutyloxy chiral tail. As shown in Table 1, pure DOBAMBC exhibits a smectic C* phase with a ferroelectric polarization of -3 nC/cm.sup.2.
There are a number of reports of compounds containing phenylbenzoate, biphenylbenzoate, tolane, diphenyldiacetylene and related cores coupled to 1-methylalkyloxy or lactate chiral tail units which possess monotropic smectic C* phases affording useful switching properties in the Clark-Lagerwaal SSFLC light valve, or which can be employed as FLC dopants to induce high polarization, fast switching speeds, high tilt angle, high birefringence, or other useful properties when combined in mixtures with FLC host materials.
The following are exemplary reports of such FLC compounds:
Furukawa, K. et al. (1988) Ferroelectrics 85:451-459 refers to chiral smectic C compounds having an ester group in the core and an optically active tail group, either alkoxy or alkoxy carbonyl, with an electronegative substituent, either halogen or cyano group, ortho to the chiral tail, for example: ##STR5## where m=2, X=H, Halogen or CN.
Walba, et. al. (1991) Ferroelectrics 113:21-36 and Walba and Otterholm, U.S. Pat. No. 5,145,601 refers to FLC components possessing the 1-methylheptyloxy chiral tail unit in combination with pyridine and pyridine-N-oxide core units, where the nitrogen atom of the pyridine ring is adjacent to the point of attachment of the chiral tail, with formula: ##STR6## where X=an electron lone pair or oxygen.
It has been demonstrated in Walba, et. al., (1989) J. Am. Chem. Soc. 111:8273-8274 and U.S. Pat. No. 5,138,010 that for some side-chain ferroelectric liquid crystal polymers composed of a polymer backbone substituted with mesogenic units wherein the achiral tail provides the connection between the polymer backbone and the mesogenic units, the mesogenic unit in the polymer imparts ferroelectric ppolarization similarly to the low molar mass mesogen itself, though the switching speeds and alignment properties of such polymers are different than the low molar mass mesogens, the switching speeds generally being slower. An exemplary FLC side chain polymer has formula: ##STR7##
Kapitza et al., (1990) Adv. Mater. 2:539-543 have disclosed side chain FLC siloxane polymers and copolymers of formula: ##STR8##
While a number of FLCs (both pure compounds and mixtures) useful in the Clark-Lagerwaal device geometry and other related devices involving large nuclear motions in response to applied fields have been reported, there has been very little work aimed at creating FLCs for electronic NLO applications. Indeed, it has been commonly known in the art of NLO materials design that FLCs are not useful in applications where the medium must respond strongly and quickly to applied fields (either respond to DC or low frequency fields in times less than 10 nsec, or respond to AC fields with frequencies larger than 100 MHz). Such applications require response with small or no nuclear motions. For example, SHG requires response of the material to optical frequency AC applied fields, at which frequencies the molecular nuclei cannot respond.